The 2D shape structure dataset: A user annotated open access - - PowerPoint PPT Presentation

▶
the 2d shape structure dataset a user annotated open
SMART_READER_LITE
LIVE PREVIEW

The 2D shape structure dataset: A user annotated open access - - PowerPoint PPT Presentation

Dec 04, 2023 172 likes •402 views

The 2D shape structure dataset: A user annotated open access database A. Carlier, G. Morin K. Leonard S. Hahmann M. Collins IRIT INP Toulouse CSUCI INRIA Grenoble GISHWHES 24-06-2016 SMI'16 1 Motivation [Liu et al.] Convex shape

slide-1
SLIDE 1

24-06-2016 SMI'16 1

The 2D shape structure dataset: A user annotated open access database

  • A. Carlier, G. Morin

IRIT – INP Toulouse

  • K. Leonard

CSUCI

  • S. Hahmann

INRIA Grenoble

  • M. Collins

GISHWHES

slide-2
SLIDE 2

24-06-2016 SMI'16 2

Motivation

Ground Truth ?

[Liu et al.] Convex shape decomposition. In: IEEE conference

  • n computer vision and pattern recognition (CVPR); 2010.

[Lien et al.] Approximate convex decomposition of polygons. In: Comput Geom, 35 (1–2) (2006), pp. 100–123 [Luo et al.] A computational model of the short-cut rule for 2D shape decomposition. In: IEEE Trans Image Process, 24 (1) (2015), pp. 273–283

slide-3
SLIDE 3

24-06-2016 SMI'16 3

Shape structure

Main Shape Details Parts

slide-4
SLIDE 4

24-06-2016 SMI'16 4

Existing datasets

[Jiang et al.] Toward perception-based shape decomposition. In: Computer vision,

  • ACCV. Springer; 2012. p. 188–201.

20 categories 10 shapes per category 12 human annotations per shape Segmentation, not hierarchy

slide-5
SLIDE 5

24-06-2016 SMI'16 5

Existing datasets

[De Winter et al.] Segmentation of object outlines into parts: a large-scale integrative study. In: Cognition, 99 (3) (2006), pp. 275–325

88 shapes 201 users 122 human annotations per shape in average Cuts were drawn, but no hierarchy was given

slide-6
SLIDE 6

24-06-2016 SMI'16 6

Setup

  • 1253 shapes, 70 categories
  • All shapes are from the MPEG7 database, plus a few artificial

shapes

  • Sub-sampling the boundary curves
  • Delaunay triangulation (~100 triangles / shape)
slide-7
SLIDE 7

24-06-2016 SMI'16 7

Setup

slide-8
SLIDE 8

24-06-2016 SMI'16 8

Setup

  • 15,000 participants, grouped in teams of 5
  • Our task is one item, out of 212 items
  • Annotate 20 shapes
slide-9
SLIDE 9

24-06-2016 SMI'16 9

Setup

  • Tutorial
  • 4 Gold Standard shapes
slide-10
SLIDE 10

24-06-2016 SMI'16 10

Collected data

  • 2,861 teams started the task, 1,877 completed

it

  • 41,953 annotated shapes
  • At least 24 annotations per shape
slide-11
SLIDE 11

24-06-2016 SMI'16 11

Collected data

slide-12
SLIDE 12

24-06-2016 SMI'16 12

Collected data

slide-13
SLIDE 13

24-06-2016 SMI'16 13

Collected data

slide-14
SLIDE 14

24-06-2016 SMI'16 14

Errors

Random mistakes Bad understanding of task Spamming

Oleson, D., Sorokin, A., Laughlin, G. P., Hester, V., Le, J., & Biewald, L. (2011). Programmatic Gold: Targeted and Scalable Quality Assurance in Crowdsourcing. Human computation, 11(11).

slide-15
SLIDE 15

24-06-2016 SMI'16 15

Filtering users

Distance Quality of user u

slide-16
SLIDE 16

24-06-2016 SMI'16 16

Worst users

slide-17
SLIDE 17

24-06-2016 SMI'16 17

Majority Vote

slide-18
SLIDE 18

24-06-2016 SMI'16 18

Is the majority vote satisfying?

slide-19
SLIDE 19

24-06-2016 SMI'16 19

Is the majority vote satisfying?

slide-20
SLIDE 20

24-06-2016 SMI'16 20

Spectral clustering

Affinity Matrix

Ng, A. Y., Jordan, M. I., & Weiss, Y. (2002). On spectral clustering: Analysis and an algorithm. Advances in neural information processing systems, 2, 849-856.

slide-21
SLIDE 21

24-06-2016 SMI'16 21

Results

slide-22
SLIDE 22

24-06-2016 SMI'16 22

Conclusion

  • Dataset is available at:

http://2dshapesstructure.github.io/index.html

  • Raw data is presented along with processed

results: majority vote, and spectral clustering

  • All related code can be found here:

https://github.com/2DShapesStructure/